Research Article pubs.acs.org/acscatalysis
Palladium Nanoparticles Encapsulated in Core−Shell Silica: A Structured Hydrogenation Catalyst with Enhanced Activity for Reduction of Oxyanion Water Pollutants Yin Wang,†,‡,∥ Jinyong Liu,†,‡ Peng Wang,§ Charles J. Werth,†,⊥ and Timothy J. Strathmann*,† †
Department of Civil and Environmental Engineering, University of Illinois at Urbana−Champaign, Urbana, Illinois 61801, United States § Water Desalination and Reuse Center, Biological and Environmental Sciences and Engineering Division, King Abdullah University of Science and Technology, Thuwal, Saudi Arabia S Supporting Information *
ABSTRACT: Noble metal nanoparticles have been applied to mediate catalytic removal of toxic oxyanions and halogenated hydrocarbons in contaminated water using H2 as a clean and sustainable reductant. However, activity loss by nanoparticle aggregation and difficulty of nanoparticle recovery are two major challenges to widespread technology adoption. Herein, we report the synthesis of a core−shell-structured catalyst with encapsulated Pd nanoparticles and its enhanced catalytic activity in reduction of bromate (BrO3−), a regulated carcinogenic oxyanion produced during drinking water disinfection process, using 1 atm H2 at room temperature. The catalyst material consists of a nonporous silica core decorated with preformed octahedral Pd nanoparticles that were further encapsulated within an ordered mesoporous silica shell (i.e., SiO2@Pd@mSiO2). Well-defined mesopores (2.3 nm) provide a physical barrier to prevent Pd nanoparticle (∼6 nm) movement, aggregation, and detachment from the support into water. Compared to freely suspended Pd nanoparticles and SiO2@Pd, encapsulation in the mesoporous silica shell significantly enhanced Pd catalytic activity (by a factor of 10) under circumneutral pH conditions that are most relevant to water purification applications. Mechanistic investigation of material surface properties combined with Langmuir−Hinshelwood modeling of kinetic data suggest that mesoporous silica shell enhances activity by promoting BrO3− adsorption near the Pd active sites. The dual function of the mesoporous shell, enhancing Pd catalyst activity and preventing aggregation of active nanoparticles, suggests a promising general strategy of using metal nanoparticle catalysts for water purification and related aqueous-phase applications. KEYWORDS: palladium nanoparticle, mesoporous silica, core−shell, bromate, oxyanion, water purification ceutical micropollutant diatrizoate,18 all of which are highly resistant to treatment by conventional drinking water treatment technologies. However, a major challenge to the application of Pd NPs for water purification is the susceptibility of NPs that lack surface stabilizers to aggregate in water to form large bulk Pd precipitates, and particle aggregation negatively affects reactivity of the NPs.19−21 In addition, Pd NPs are difficult to recover from water following application, because of their small size, and there are growing concerns about the release of such NPs into aquatic environments and associated ecological and public health risks.22 A variety of immobilization strategies are being investigated to overcome limitations and concerns associated with applying Pd NPs in aquatic systems. Core−shell support materials decorated with noble-metal NPs have drawn increased attention in heterogeneous catalysis, because of their unique
1. INTRODUCTION Palladium-based materials have wide application in the field of catalysis, because of their ability to facilitate a variety of reactions.1−3 One emerging application is to use Pd-based catalysts in water purification and remediation processes for the reductive transformation of recalcitrants and emerging classes of aquatic contaminants, including halogenated organics, oxyanions, nitrosamines, and pharmaceutical and personal care products (PPCPs).3−8 Recent developments in material science and technology have enabled the synthesis of Pd nanoparticles (NPs) with controllable shapes and sizes.9−13 Pd NPs and Pd-based bimetallic NPs have been applied in catalyzing various aqueous reactions,14,15 and have shown great potential as catalysts for water treatment and purification application, because of their unique properties, such as high surface area to volume ratio and quantum size effects.16−18 Shuai et al. synthesized Pd NPs with various shapes and sizes and demonstrated their ability to catalytically reduce the oxyanion nitrite (NO2−), the disinfection byproduct Nnitrosodimethylamine (NDMA), and the halogenated pharma© XXXX American Chemical Society
Received: July 8, 2014 Revised: September 1, 2014
3551
dx.doi.org/10.1021/cs500971r | ACS Catal. 2014, 4, 3551−3559
ACS Catalysis
Research Article
Scheme 1. Preparation Procedure of the Core−Shell-Structured Silica Materials with Encapsulated Pd NPs
electronic features that affect catalytic activity, lack of aggregation in aqueous solution, and prevention of sintering at elevated temperatures (e.g., >350 °C).23−25 Silica-based core−shell materials have received special attention, because of their high thermal stability and tolerance of acidic conditions.26,27 Successful synthesis of several types of core−shell silica materials with immobilized noble-metal NPs has been reported.28−31 Joo et al. developed a core−shell-structured nanocatalyst consisting of a single platinum (Pt) nanoparticle core coated with a mesoporous silica shell (i.e., Pt@mSiO2).23 They found that the presence of the mesoporous shell prevented the aggregation and Pt NP sintering at high temperature, thereby enhancing catalytic CO oxidation. Similarly, synthesis of a Pd-based core−shell-structured silica was reported with a single Pd NP core encapsulated in a mesoporous silica shell.32 Although these materials have promise in gas-phase applications, their small size (i.e., 15 nm). In designing core−shell-structured supports for water purification applications, ideally the mesoporous shell will be highly porous, to allow the transport of reactants to encapsulated metal NPs, but individual pore sizes will be narrower than the size of metal NPs, to effectively prevent NP aggregation or release from the support material. The primary objective of this contribution was to develop an advanced silica-based structure to enhance the suitability of Pd NPs in environmental catalysis application. We report the design, synthesis, and aqueous reactivity of a silica-based core− shell-structured material with encapsulation of shaped Pd NPs. A nonporous silica core was decorated with preformed octahedral Pd NPs that were further encapsulated within a surfactant-templated mesoporous silica shell developed through
a NaOH-mediated hydrolysis approach (i.e., SiO2@Pd@ mSiO2). The well-designed pore structure has a small pore size and narrow size distribution, and it makes the material highly suitable as a model catalyst for probing aqueous reactions. The catalyst was applied to remove bromate (BrO3−) as a model drinking water contaminant. Bromate is an EPA-regulated byproduct of drinking water disinfection processes, because of its nephrotoxicity and potential carcinogenicity.35−37 To the best of our knowledge, this is the first report of the construction of SiO2@Pd@mSiO2 with a well-defined pore structure with Pd NPs constrained to the interface between the silica core and the mesoporous shell. More importantly, this work demonstrates, for the first time, that the presence of mesoporous silica shell enhances Pd catalyst reactivity with contaminants under aqueous conditions relevant to water purification applications.
2. RESULTS AND DISCUSSION 2.1. Synthesis and Characterization of SiO2@Pd@ mSiO2. SiO2@Pd@mSiO2 microspheres were prepared in three steps (Scheme 1): (1) separate synthesis of shapecontrolled Pd NPs and surface-functionalized nonporous silica microspheres, (2) immobilization of the preformed Pd NPs on the surface of the silica microspheres (SiO2@Pd), and (3) growth of a surfactant-templated mesoporous silica shell on the SiO 2@Pd microspheres. A combination of techniques, described individually in detail in the Experimental Section, were used in the synthesis process, including sol−gel processing, interfacial deposition, and surfactant templating. Pd NPs were synthesized by reducing Na2PdCl4 with citric acid, as reported previously.13 Polyvinylpyrrolidone (PVP) was added to serve as a stabilizer to control the growth and prevent the aggregation of Pd NPs. As observed by transmission electron microscopy (TEM), the Pd NPs were mainly in an octahedral shape (>75%) with an edge length of ∼6 nm (see Figures 1A and 1B). The X-ray diffraction (XRD) pattern of the Pd NPs matched well with the face-centered cubic (fcc) Pd(0) crystal reference pattern (Figure 2). No crystalline phases other than Pd(0) were observed from XRD. The average Pd NP size was calculated as 6.2 nm from peak widths, according to the Debye−Scherrer equation, which is consistent with the observation from TEM. Silica microspheres were synthesized using the Stöber method.38 As shown in the scanning electron microscopy (SEM) image (Figure 3A), monodisperse silica microspheres with a diameter of ∼400 nm were obtained. Ammonia was added to provide basic conditions to facilitate tetraethylortho3552
dx.doi.org/10.1021/cs500971r | ACS Catal. 2014, 4, 3551−3559
ACS Catalysis
Research Article
Figure 3. SEM images of (A) nonporous SiO2 microspheres, (B) −NH2 functionalized SiO2 microspheres, (C, D) SiO2@Pd microspheres, and (E, F) SiO2@Pd@mSiO2 microspheres. Figure 1. TEM images of (A, B) Pd NPs, (C, D) SiO2@Pd microspheres, and (E, F) SiO2@Pd@mSiO2 microspheres.
Based on the Debye−Scherrer equation, the Pd particle size on SiO2@Pd was calculated as 6.5 nm, which is in close agreement with measurements of the free Pd NPs. Results suggest that both the size and the crystal structure of the Pd NPs were preserved during immobilization on the silica core. For the SiO2@Pd, the Pd NPs were strongly attached to the surface of the −NH2-functionalized silica sphere. In addition to the interaction between Pd NP and −NH2 group, PVP that is coated on the Pd NP surface might also facilitate the immobilization of Pd NPs by forming hydrogen bonds with −NH2 groups on the silica surface or by modifying surface charge to enhance the electrostatic interactions.42 Attempts to immobilize the Pd NPs on unfunctionalized and −SHfunctionalized silica microspheres, which were previously reported to promote Au NP immobilization, 29,43 were unsuccessful. The mass loading of Pd NPs immobilized on the silica surface can be tuned by adjusting the relative amounts of Pd NPs and −NH2-functionalized silica microspheres (see Figure S1 in the Supporting Information (SI)). Almost all the Pd NPs in aqueous suspension were immobilized when their amount was 15 nm) than that reported here. It is worth noting that the pore size obtained here is significantly smaller than the size of the preformed Pd NPs, thereby providing a physical barrier to prevent release of the NPs into solution. In addition, it is worth mentioning that the surfactant templating method may also result in a highly ordered pore orientation that is perpendicular to the core, since this preferred alignment of surfactant micelles and TEOS minimizes the overall interfacial energy of the system.33,46,47 The thickness of the mesoporous shell can be controlled by adjusting the amounts of TEOS and CTAB. For example, increasing the TEOS concentration 2.4-fold during synthesis resulted in an increase of the mesoporous shell thickness from 60 nm to 120 nm (see Figures S3 and S4 in the SI). Interestingly, further increasing the TEOS concentration by another 50% resulted in a nonuniform coating of mesoporous silica shell, and SiO2@Pd@mSiO2 microspheres with multiple sizes were obtained (see Figure S4 in the SI). 2.2. Catalytic Reduction of Bromate. The activity of SiO2@Pd@mSiO2 in catalyzing the reduction of bromate (BrO3−) was measured and compared with that of free Pd NPs and SiO2@Pd in H2-saturated water at pH 7 and ambient temperature (20 ± 1 °C). The overall reaction can be described as
combined with the use of water as the solvent yielded a more uniform distribution of Pd NPs on silica microsphere surfaces. A surfactant templating method was used to grow mesoporous silica shells on the SiO2@Pd microspheres. TEOS was hydrolyzed in the presence of SiO2@Pd and hexadecyltrimethylammonium bromide (CTAB), a structure directing agent that promotes formation of a mesostructured CTAB/silica shell on the surface of the SiO2@Pd microspheres.44 The CTAB then was removed by extraction in a mixture of ethanol and concentrated HCl to obtain SiO2@Pd@ mSiO2 microspheres.45 The thickness of the mesoporous shell was ∼60 nm and the Pd content of the SiO2@Pd@mSiO2 was determined to be 1.3 wt %. Instead of NH3·H2O,33 dilute NaOH (1.7 mM) was used as a base during the surfactanttemplating procedure to catalyze TEOS hydrolysis and preserve the sphere shape of the mesoporous shell. With the use of NaOH, sandwich-structured SiO2@Pd@mSiO2 microspheres were obtained where Pd NPs were well-constrained at the interface between the silica core and shell materials (Figure 1E and 1F). In contrast, when using NH3·H2O as a base the majority of the Pd NPs were liberated from the surface of the inner silica core and distributed randomly throughout the emplaced mesoporous silica shell (see Figure S2 in the SI). The loss of Pd NPs from the core surface may be due to competing Pd interactions with −NH2 groups on the silica surface and dissolved NH3. Using NaOH to form the mesoporous shell also resulted in uniform silica spheres (see Figure 1E). Joo et al. also reported uniform core−shell-structured Pt@mSiO2 spheres when adding diluted NaOH to catalyze surfactant-templated TEOS hydrolysis in the presence of free Pt NPs.23 It is worth noting that previous studies report that higher NaOH concentrations (>0.5 M) may react directly with silica, etching the mesoporous silica structure and changing the pore size.28,34 The porous nature of the SiO2@Pd@mSiO2 microspheres was confirmed by measuring N2 adsorption−desorption isotherms (Figure 4). The Brunauer−Emmett−Teller (BET)
Pd
BrO−3 + 3H 2 → Br − + 3H 2O
(1)
For all three materials, BrO3− consumption was accompanied by Br − production, and the sum of BrO 3 − and Br − concentrations remained very close to the initial BrO3− concentration (see Figure S5 in the SI), confirming that the observed BrO3− loss from solution resulted from reduction of the oxyanion rather than adsorption to the high surface area material. None of the oxyanion intermediates (BrO2− and BrO−) were detected, indicating that the reduction of such intermediates is much faster than that of BrO3−. Comparing the time courses for BrO3− reduction observed with the three materials under neutral pH conditions (Figure 5A) reveals that the reaction with SiO2@Pd@mSiO2 proceeded much faster than with Pd NPs and SiO2@Pd. Kinetics followed a pseudo-first-order rate law (Figure 5A), and the resulting rate constants (Pd mass normalized) were determined for all the experiments (see Table 1). Pseudo-first-order rate constants are typically used to describe the kinetics of different catalytic reduction experiments.5,8,48−51 Similar rate constants for BrO3− reduction with Pd NPs and SiO2@Pd were measured at pH 7 (Figure 5B). Thus, immobilization of Pd NPs onto silica supports did not decrease the apparent catalytic activity. Under these same pH conditions, which is highly relevant to water treatment applications, the rate constant for SiO2@Pd@mSiO2 was more than an order-of-magnitude larger than those obtained for Pd NPs and SiO2@Pd (Figure 5B). Initial turnover frequency (TOF0) values were also determined (Table 1); they followed the same trend as the pseudo-firstorder rate constants. To the best of our knowledge, this is the first report of mesoporous-shell-enhanced activity of metal NPs in catalysis applications at ambient temperature. These results indicate that, in addition to the active metal sites, the structure of the solid supports play an important role in determining the
Figure 4. N2 adsorption−desorption isotherms and pore size distribution (inset) of SiO2@Pd@mSiO2 microspheres prepared with an ∼60-nm-thick mesoporous shell.
surface area was measured to be 480 m2/g, and the total pore volume was estimated as 0.27 cm3/g. Application of the Barrett−Joyner−Halenda (BJH) model confirmed that the surfactant templating approach yielded a highly uniform pore size distribution in the mesoporous shell with an average diameter of 2.3 nm (see Figure 4). The value is similar to that reported by Deng et al., who also used CTAB to form core− shell-structured Fe3O4@SiO2−Au@SiO2 microspheres with an 3554
dx.doi.org/10.1021/cs500971r | ACS Catal. 2014, 4, 3551−3559
ACS Catalysis
Research Article
Figure 5. Reduction of 100 μM BrO3− by 2 mgPd L−1 loading of Pd NPs, SiO2@Pd, and SiO2@Pd@mSiO2 in water at pH 7 (20 ± 1 °C, PH2 = 1 atm): (A) Time courses for BrO3− reduction (dashed lines represent pseudo-first-order kinetics model fits); (B) Pd mass-normalized rate constants for BrO3− reduction (error bars represent one standard deviation determined from triplicate experiments).
Table 1. Results of Kinetics Experiments Measuring Catalytic Reduction of Bromatea catalyst Pd NP SiO2@Pd SiO2@Pd SiO2@Pd SiO2@Pd SiO2@Pd SiO2@Pd@ mSiO2 SiO2@Pd@ mSiO2 SiO2@Pd@ mSiO2 SiO2@Pd@ mSiO2 SiO2@Pd@ mSiO2
pH
rate constant, kb (L h−1 gPd−1) (±3.69) (±0.53) (±1.03) (±0.55) (±0.66) (±0.65) (±0.56)
× × × × × × ×
100 100 100 101 102 103 101
initial turnover frequency, TOF0b (h−1) × × × × × × ×
100 100 100 100 101 102 100
7 8 7 6 4 2 8
6.88 2.47 5.45 1.56 5.07 2.62 2.89
7
6.60 (±1.69) × 101
7.40 (±1.89) × 100
6
9.39 (±2.55) × 101
1.52 (±0.41) × 101
4
4.96 (±0.29) × 102
5.56 (±0.33) × 101
2
2.86 (±0.37) × 103
3.20 (±0.41) × 102
0.30 0.21 0.47 1.34 4.35 2.25 3.23
(±0.16) (±0.05) (±0.09) (±0.47) (±0.57) (±0.56) (±0.63)
Figure 6. Influence of pH on rate constants for reduction of 100 μM BrO3− by 2 mgPd L−1 loading of SiO2@Pd and SiO2@Pd@mSiO2 in water (T = 20 ± 1 °C, PH2 = 1 atm). Error bars represent one standard deviation of replicate experiments.
All experiments were conducted at [BrO3−]0 = 100 μM, T = 20 ± 1 °C, PH2 = 1 atm, and catalyst loading equivalent to yield 2 mgPd L−1. b Uncertainty represents one standard deviation derived from at least duplicate experiments. a
catalyst.49 They proposed that acidic condition promoted hydrogen bonding between ClO4− and the immobilized Re metal species, thereby accelerating oxygen atom transfer in the proposed reaction mechanism. Chen et al. found that the reduction rate of bromate increased with decreasing pH with an alumina supported Pd catalyst, and they attributed the pHdependent activity to both the redox potential of bromate reduction and the surface properties of the catalyst.53 The zeta potentials of SiO2@Pd and SiO2@Pd@mSiO2 were measured in H2-saturated aqueous solutions as a function of pH in this work (Figure 7). The surfaces of both SiO2@Pd and SiO2@ Pd@mSiO2 were less negatively charged with decreasing pH. Similar to others, we propose that less negative surface charge at lower pH increases electrostatic attractive interactions between BrO3− and the catalyst surfaces and increases the apparent reaction rate. It is worth noting that the presence of dissolved H2 markedly affects the zeta potential of Pd-based catalyst surfaces, and here the addition of H2(g) shifted the zeta potential of both SiO2@Pd and SiO2@Pd@mSiO2 to a more negative values at a given pH (see Figure S6 in the SI). Similarly, Choe et al. reported negative shifts in the surface charge of Pd/C catalysts in the presence of H2.54 The negative shift of zeta potential may be attributed to the formation of activated H species, which may adsorb on the surface of Pd or react with support surface functional groups, leading to the formation of negatively charged species.54
activity of heterogeneous catalysts. The mechanism of enhancement is examined in a subsection below. The effect of pH on reaction kinetics was determined for the two silica-supported catalysts (SiO2@Pd and SiO2@Pd@ mSiO2). BrO3− reduction rates increased with decreasing pH for both SiO2@Pd and SiO2@Pd@mSiO2 (Figure 6). SiO2@ Pd@mSiO2 was found to be much more active than SiO2@Pd for pH 6−8, but activities for the two materials converged under more acidic conditions. Compared to metal- or metaloxide-based supports, including alumina and iron, silica is not only stable under mild basic conditions (pH
